Customized classifier over common features

Abstract

A method of updating a set of classifiers includes applying a first set of classifiers to a first set of data. The method further includes requesting, from a remote device, a classifier update based on an output of the first set of classifiers or a performance measure of the application of the first set of classifiers.

What is claimed is: 
     
         1 . A method of updating a set of classifiers comprising:
 applying a first set of classifiers to a first set of data; and   requesting, from a remote device, a classifier update based at least in part on at least one of an output of the first set of classifiers or a performance measure of the application of the first set of classifiers.   
     
     
         2 . The method of  claim 1 , in which the requesting is based at least in part on context information. 
     
     
         3 . The method of  claim 1 , in which the performance measure comprises an accuracy of the classifiers, a level of agreement of multiple classifiers, or a speed of computation of the classifiers. 
     
     
         4 . The method of  claim 1 , in which the first set of classifiers and the classifier update are built on a same feature generator. 
     
     
         5 . The method of  claim 1 , in which the first set of classifiers comprises a general classifier and the classifier update comprises a specific classifier. 
     
     
         6 . The method of  claim 5 , further comprising applying the specific classifier to an object to identify a specific class of the object. 
     
     
         7 . The method of  claim 1 , in which the remote device is configured to apply the first set of classifiers. 
     
     
         8 . The method of  claim 7 , further comprising:
 computing features and transmitting the computed features to the remote device, the remote device applying the first set of classifiers to the computed features to compute a classification.   
     
     
         9 . An apparatus for updating a set of classifiers comprising:
 a memory; and   at least one processor coupled to the memory, the at least one processor being configured:
 to apply a first set of classifiers to a first set of data; and 
 to request, from a remote device, a classifier update based at least in part on at least one of an output of the first set of classifiers or a performance measure of the application of the first set of classifiers. 
   
     
     
         10 . The apparatus of  claim 9 , in which the at least one processor is further configured to request the classifier update based at least in part on context information. 
     
     
         11 . The apparatus of  claim 9 , in which the performance measure comprises an accuracy of the classifiers, a level of agreement of multiple classifiers, or a speed of computation of the classifiers. 
     
     
         12 . The apparatus of  claim 9 , in which the first set of classifiers and the classifier update are built on a same feature generator. 
     
     
         13 . The apparatus of  claim 9 , in which the first set of classifiers comprises a general classifier and the classifier update comprises a specific classifier. 
     
     
         14 . The apparatus of  claim 13 , in which the at least one processor is further configured to apply the specific classifier to an object to identify a specific class of the object. 
     
     
         15 . The apparatus of  claim 9 , in which the remote device is configured to apply the first set of classifiers. 
     
     
         16 . The apparatus of  claim 15 , in which the at least one processor is further configured:
 to compute features and transmit the computed features to the remote device, the remote device applying the first set of classifiers to the computed features to compute a classification.   
     
     
         17 . An apparatus for updating a set of classifiers comprising:
 means for applying a first set of classifiers to a first set of data; and   means for requesting, from a remote device, a classifier update based at least in part on at least one of an output of the first set of classifiers or a performance measure of the application of the first set of classifiers.   
     
     
         18 . A computer program product for updating a set of classifier comprising:
 a non-transitory computer readable medium having encoded thereon program code, the program code comprising:
 program code to apply a first set of classifiers to a first set of data; and 
 program code to request, from a remote device, a classifier update based at least in part on at least one of an output of the first set of classifiers or a performance measure of the application of the first set of classifiers.